Classifying Unordered Feature Sets with Convolutional Deep Averaging Networks
This addresses a data structure issue in machine learning for domains with unordered sets, but it is incremental as it builds on existing permutation-invariant methods.
The paper tackles the problem of classifying variable-size, unordered feature sets, which traditional neural networks handle poorly, by proposing convolutional deep averaging networks (CDANs) that are efficient, permutation-invariant, and handle arbitrary sizes, with experiments showing advantages over linear embeddings and other architectures.
Unordered feature sets are a nonstandard data structure that traditional neural networks are incapable of addressing in a principled manner. Providing a concatenation of features in an arbitrary order may lead to the learning of spurious patterns or biases that do not actually exist. Another complication is introduced if the number of features varies between each set. We propose convolutional deep averaging networks (CDANs) for classifying and learning representations of datasets whose instances comprise variable-size, unordered feature sets. CDANs are efficient, permutation-invariant, and capable of accepting sets of arbitrary size. We emphasize the importance of nonlinear feature embeddings for obtaining effective CDAN classifiers and illustrate their advantages in experiments versus linear embeddings and alternative permutation-invariant and -equivariant architectures.